The agreements, the channels, and why the vehicle you buy through decides more than the discount you negotiate. Three knowledge checks along the way, and 1 clip from a senior cloud advisor.
This is a taught session, not a talking head. The instructor works through analyst grade slides, and three times the video stops on a question with four options on screen. Pause, commit to an answer, and the next slide explains which option is right and why each of the others is wrong. Once in the session the frame splits and a senior cloud advisor gives the view from inside real Oracle negotiations, and the instructor picks the clip apart when the slides return.
The full narration of this session, section by section, for reading and reference. Guest analyst clips are marked.
Welcome to session one of forty. Before we go anywhere near a price, we need to be clear about something that decides more money than any discount you will ever negotiate: the vehicle you buy through. An Enterprise Agreement, a Microsoft Customer Agreement, and a CSP contract are three different machines, and the same estate priced through each of them lands in three different places. Over the next forty sessions we will cover all of it, the agreements, the true ups, the Microsoft 365 tiers from E1 through E7, Copilot and the consumption layer underneath it, the renewal, and a full worked negotiation at the end. Today is the map: who is actually in the transaction, what the three vehicles are, the five things that set your price, and what changed under buyers in the last two years. Let's start with who gets paid.
Five takeaways from today. One, the map: Microsoft's field team, the licensing solution provider who transacts your agreement, the CSP partner, the deal desk nobody ever meets, and where each one's money comes from. Two, the vehicles: the Enterprise Agreement, the Microsoft Customer Agreement, and CSP, three machines that price the same estate differently, and the decision module one exists to make properly. Three, the price shape: five inputs that set your number, ranked by how much they actually move, and the headline discount comes last on that list, deliberately. Four, the 2026 pressure: two price waves inside twenty four months, compressed volume tiers, and the push away from the EA, which is the environment your next renewal happens in. And five, how this course works, because forty sessions is a commitment and you should know what you are getting. One promise before we start: everything here is buyer side. Nothing in this course is designed to help you feel good about spending more.
Who is actually in the transaction. The Microsoft field team owns your account, the renewal, and a quota, and their money comes from your growth: new workloads, tier moves, and right now, Copilot above almost everything else. The licensing solution provider transacts the Enterprise Agreement, handles the paperwork and the true ups, and takes a margin that lives inside the pricing you see, which is worth knowing because that margin is negotiable and most buyers never ask. The CSP partner resells subscriptions, bills you directly, and provides your first line support, earning partner margin plus whatever services they attach. The deal desk approves discounts and non standard terms, never appears in your meetings, and is the party that actually says yes or no to the clause you are asking for. And then there is you, signing either a three year enterprise wide commitment or an evergreen one. Two facts to carry out of this slide. The person sitting in front of you rarely holds the discount authority, so escalation is not rudeness, it is arithmetic. And every route to identical software carries somebody's margin, so the question is never whether there is margin, it is whose, and how much of it is negotiable.
The three vehicles, in outline, because the rest of module one takes them one at a time. The Enterprise Agreement: three year term, enterprise wide commitment, annual true ups that bill your growth at contracted rates, and per unit prices locked for the term. It is the strongest vehicle on price for a large stable base, and the weakest when your base shrinks, because the true up counts up and never down. The Microsoft Customer Agreement: evergreen, no fixed end date, sold in Enterprise, Online, and Partner forms, and it is Microsoft's preferred direction of travel. Evergreen flexibility cuts both ways, and this is the sentence to remember: nothing expires, which also means nothing is protected unless you engineered the protection yourself. And CSP: bought through a partner who bills and supports you, where monthly terms let you reduce your count and annual terms do not, strongest for a smaller or more volatile base, and the only vehicle where the quality of a partner is part of the product you are buying. Now the number in the note, because it anchors the whole comparison: across the cost comparisons we have run, the break even between CSP and the EA sits near a stable base of about two thousand four hundred users. Above it the EA rate typically wins by six to fourteen percent. Below it CSP undercuts by five to twelve, because you stop paying for idle committed seats. First knowledge check.
First check. Two companies license an identical Microsoft 365 estate: three thousand seats, same tier mix, same country. One pays materially less per seat. The most likely explanation: A, one negotiated a better discount percentage. B, one bought through a different vehicle with a different term, commitment, and price protection structure, and the vehicle difference outweighs the discount difference. C, Microsoft charges different list prices by customer. Or D, one is larger and gets automatic volume pricing. Pause here. What did each of those two buyers commit to, and for how long?
The answer is B, and it is the thesis of the entire course in one question. Discount percentages are the visible variable, the one that gets reported to the board, and they are rarely the decisive one. The vehicle decides the term length, whether your per unit price is held for three years or drifts every year, whether your count can ever go down, and what happens at the anniversary, and those structural differences routinely swamp a few points of discount. A is real but secondary, and notice how the two interact: a great discount inside a structure that cannot shed seats is worse than a mediocre discount inside one that can. C is simply wrong, list price is public and uniform, what differs is what you negotiated against it and what you committed to in exchange. And D confuses size with structure: size determines which tier you qualify for, the vehicle determines what that tier is worth over three years. So the opening discipline of this course: negotiate the frame before you negotiate the number inside it. Everything in module one is about the frame.
What actually shapes your price, five inputs, ranked by how much they move. One, the vehicle and the term, which we have just met and which sets the boundaries everything else operates inside. Two, the mix: how many people sit on each tier, and this one is worth internalising early, moving two thousand users from E5 to E3 changes more money than any discount conversation you will ever have, and unlike the discount it is entirely within your control. Three, the commitment: what you promise, in seats, in Azure spend, in Copilot quantities, and understand the trade, commitment buys discount and sells flexibility, and the exchange rate between those two is negotiable. Four, price protection: whether your rate survives the term and what happens at renewal, and in an evergreen agreement this exists only because you wrote it in. And five, last and deliberately so, the discount. It is the number everybody reports upward and the one that moves least once the four above it are settled. Our guest analyst spends his working life on these deals, and he has a story about exactly this ranking.
Guest analyst The deal that taught me to distrust the headline rate was a manufacturing group, about four thousand seats, renewing their Enterprise Agreement. Their sourcing lead had done genuinely good work: he had pushed the account team from an eleven percent discount to nineteen, and he was rightly pleased. Eight points is real money. Then we looked at the mix, which nobody had touched. Three thousand one hundred of those four thousand seats were on E5. When we pulled the admin centre data and matched it against what people actually used, the security and compliance features that justify E5 were being used by about nine hundred people. The rest had it because a previous rollout had found it simpler to give everybody the same thing. Right sizing that mix, moving roughly two thousand users down to E3 and keeping E5 where the features were genuinely used, was worth about four times the entire eight point discount win. And here is the part that stuck with me: when we raised it, the account team was perfectly happy to discuss the discount all day and noticeably less keen to discuss the mix. Of course they were. The discount comes out of their margin. The mix comes out of their revenue. So my rule, and I give it to every client on day one: bring your own usage data to the table before you talk about price, because the vendor will happily negotiate the number that costs them least, and the mix is not it.
Eight points of discount, and the untouched mix was worth four times as much. The vendor will negotiate the number that costs them least, and the mix is not it. Hold that, because module three spends five sessions on the tiers and the mix model. Second check.
Check two. An eighteen hundred seat company with seasonal contractors and flat growth is offered an Enterprise Agreement at a six percent better per seat rate than its current CSP arrangement. The right analysis: A, take the EA, a better per seat rate on the same software is strictly better. B, stay on CSP on principle, partners always beat direct. C, price the flexibility and not just the rate: below roughly two thousand four hundred stable users the EA's committed seats often cost more than the rate saves, because seasonal peaks become permanent commitments for three years. Or D, split the estate randomly across both to hedge. Pause here. What happens to a contractor seat in month four of an Enterprise Agreement, and what happens to that same seat on monthly CSP?
The answer is C. The mechanism is the true up: an EA takes your growth up at the anniversary and does not take it back down inside the term, so a seasonal peak of contractors becomes three years of committed seats, whereas monthly CSP lets the count fall when those people leave. Six percent off a rate you pay on seats you no longer use is not six percent off anything. And the evidence is uncomfortable: across the comparisons we have run, buyers who compared on rate alone picked the wrong vehicle in close to half of cases, and this is exactly the error that produced that number. A treats a rate as if it were the whole cost. B replaces analysis with loyalty, and I want to be fair to the EA here, above the break even it genuinely does win, by six to fourteen percent, which is why session five builds a proper decision framework rather than a preference. D is not a strategy as written, but notice what sits just next to it: a deliberate split, your stable base on one vehicle and your volatile population on another, is often the right answer, and that is a real portfolio decision rather than a hedge.
What changed under buyers recently, because the structures held while the pricing environment moved twice. First, the price waves: Microsoft lifted pricing twice inside twenty four months, compounding roughly eleven to nineteen percent against the prior cycle, and on top of that the volume tier discounts on the enterprise products compressed through 2024 and 2025, adding another four to nine percent. Read those together and a renewal that looks like a fifteen percent increase may be a considerably larger one against what you would have paid on the old tiers. Second, the MCA push: the transition proposals we benchmarked priced eight to seventeen percent above the equivalent EA renewal on like for like scope for most large estates. The MCA was genuinely the right answer for highly variable consumption profiles; everywhere else the modernisation story carried a premium, and it was not always disclosed as one. And third, the good news, which is the reason this course exists: preparation still pays, and it pays more than it used to. The median final discount landed five to nine points above the band the account team initially flagged when the buyer arrived a hundred and eighty days out with a credible alternative, and renewal advisory savings ran twelve to twenty eight percent across the benchmark set. The market got harder, the preparation premium got bigger, and the buyers who lost most were the ones who ran the renewal the way they always had.
How the course is built, briefly, so you know the shape of the commitment. Eight modules of five sessions. Module one, where we are, is the landscape: the EA, the MCA, CSP, and choosing between them. Module two is EA mechanics: enrollments and price levels, the annual order and the true up, the metrics, Software Assurance, and reading a quote. Module three is the Microsoft 365 tiers: E1, E3, E5, E7, and the mix model that our guest analyst just made the case for. Module four is Copilot: the subscription, the consumption layer, Copilot cowork, the negotiation, and proving value. Module five is the rest of the estate under the agreement, Azure commitments, Windows and servers, Dynamics, Power Platform, and security. Module six is the renewal, end to end. Module seven is compliance, governance, and FinOps. And module eight is advanced situations and a capstone negotiation with everything on one table. Every session runs the same shape: the material, three knowledge checks with worked answers, about an hour of homework that produces something you can actually use, and five further reads. There is a module test after every five sessions, and a certification exam at the end for anyone who wants the credential.
Last check. Your account team tells you the Enterprise Agreement is being retired for your segment, and offers an MCA transition at a headline discount above your current EA. What is the first thing to establish? A, whether the discount percentage beats the current one. B, whether the like for like total cost over the same horizon is genuinely lower, since benchmarked MCA transitions priced eight to seventeen percent above the equivalent EA renewal, and what price protection replaces the fixed term hold you are giving up. C, which partner would transact it. Or D, how quickly the migration can be completed. Pause here. An evergreen agreement has no expiry date. What else in it has no expiry?
The answer is B, and the trick in the question is the word evergreen, which sounds like freedom and functions like exposure. No expiry means no fixed price hold, and across the transitions we advised, buyers met five to fifteen percent annual drift that nobody had modelled, and discovered that their negotiated EA discount levels did not automatically carry forward. Both of those are recoverable, but only if you know to ask before signing rather than after. So the two questions are: what is the total over a comparable horizon, not the year one headline, and what protection am I writing to replace the one the three year term used to give me for free. A compares percentages across different structures, which is the specific error this session exists to kill. C and D are execution questions, and they matter, but after the decision rather than instead of it. Session three does the MCA properly, and session thirty seven does the migration mechanics, including the timing, because the answer is not always no. Sometimes the MCA genuinely is right. It just should not be right because someone told you the EA was being retired and you did not check the arithmetic.
Three habits that run through all forty sessions, and if you take nothing else from today, take these. One, own your own numbers: seats assigned against seats actually used, tier by tier, from your own admin centre, refreshed on a schedule. The party with better data sets the terms of the conversation, and it should never be the party selling to you. Two, negotiate structure before price: vehicle, term, protections, and reduction rights first, discount last, because once the price is agreed the deal is closed in the seller's mind and every clause you raise afterwards gets bought back at a premium. And three, start early enough that you can wait. The benchmark is unambiguous: arriving a hundred and eighty days out with a credible alternative was worth five to nine points on the final discount. Time is the one piece of leverage available to every buyer regardless of size, and it is the one most estates spend without noticing. Next session we open the Enterprise Agreement itself: the three year term, the enterprise wide commitment, the two product tiers with their separate discount frameworks, and the true up that decides what your growth actually costs.
Session one, three sentences. One: three vehicles price the same estate differently, and the vehicle decides the term, the protection, and whether your count can ever fall, which routinely outweighs a few points of discount. Two: five inputs set the price, the vehicle, the mix, the commitment, the protection, and the discount, in that order of impact, and the industry reports them in exactly the reverse order. Three: the pricing environment got harder, two waves inside twenty four months plus tier compression, while the preparation premium grew, so the buyer who starts a hundred and eighty days out with their own data is the one who still does well. Next week, the Enterprise Agreement in full. See you there.
Homework, about an hour, and it is deliberately basic because everything later depends on it. One, name the vehicle: EA, MCA, CSP, or a mix, and if it is a mix, write down which populations sit where and whether anyone actually chose that or it simply accumulated. Two, find the dates: term end, anniversary, and any notice window, each in the calendar with an owner, at twelve months and at ninety days. Three, count the seats: from the admin centre, licenses assigned per tier against people who signed in during the last thirty days, and be prepared for that gap to be larger than you expect, because it usually is. That one table reappears in most of the next thirty nine sessions. Four, locate the paper: the signed agreement, the enrollments, the current price sheet, and if nobody can produce those this week, that is finding number one and it is more common than you would think. And five, ask the provenance question: who last negotiated this, when, and against what alternative, because the answer tells you roughly how much room is still on the table before you do any analysis at all.
Five reads before next session, all free on redress compliance dot com. First, the Microsoft licensing guide, the whole landscape in reference form and updated for the current cycle. Second, the Enterprise Agreement pillar, which carries the benchmark data behind today's numbers, the price waves, the tier compression, and the preparation premium. Third, the Microsoft Customer Agreement explained, the evergreen agreement and what buyers actually lost in transition. Fourth, CSP versus EA side by side, the break even analysis dimension by dimension, worth reading before session four. And fifth, the Microsoft cloud agreements and subscriptions playbook, which sits underneath the whole of module one. That is session one. Three vehicles, five inputs, and one discipline: the frame before the number. Next week, the Enterprise Agreement. See you there.